Thread by @ArunKrishnan_

A recent paper titled "Sparks of Artificial General Intelligence: Early experiments with GPT-4", published by Microsoft Research makes for very interesting reading. I have gone through the entire report and it IS TL;DR. Here is my summary of the same #GPT4 #AGI #LLMs 1/n

But before that, let's take a brief detour into what intelligence actually means. One of the definitions is that intelligence is a multifaceted and complex cognitive ability that involves the capacity to understand, learn, reason, solve problems, adapt to new situations, 2/n

think critically, and apply knowledge to different domains. But what does it mean to say that an artificial intelligence system is intelligent?
From the paper:
3/n

There are multiple other definitions of AGI:
Goal-oriented - Measures an agent's ability to achieve goals in a wide range of environments. However, it does not take into account systems that can carry out complex tasks without being motivated by a goal.
4/n

Skill-acquisition efficiency: Emphasises learning from experience.
Do anything a human can: However, there is no single, standard definition of human intelligence given the wide diversity.

5/n

In the paper, the authors propose an approach to study whether GPT4 is making progress towards AGI by sticking closer to traditional psychology. With that aim, they generated tasks and questions that were meant to push GPT4 beyond mere memorization and looked at testing GPT-4 6/n

across the following areas:
1. Its mastery of natural language by asking it to translate not just between languages but across tone, content, style and domain and to observe if it can manipulate complex concepts. Spoiler alert: It can!

7/n

2. Coding and Mathematics and its ability across tests
3. Testing its ability to plan as well as to learn from experience by having it play games and interact with tools.

8/n

4. Testing whether it can understand humans and making itself understandable by humans, i.e addressing the problem of explainability.

Having done this, as the authors say

9/n

Some interesting findings:

A key measure of intelligence is the ability to synthesize information from different domains or modalities and the capacity to apply knowledge and skills across different contexts or disciplines.

10/n

GPT-4 did some remarkable things like:

Producing javascript code which generates random images in the style of the painter Kandinsky
11/n

Find a proof that there are infinitely many prime numbers in the literary style of Shakespeare
Combining knowledge of history and physics by asking it to write a supporting letter fr Electron as a US presidential candidate, written by Mahatma Gandhi and addressed to his wife 12/n

Produce python code for a program that takes as an input a patient’s age, sex, weight, height and blood test results vector and indicates if the person is at increased risk for diabetes
There are numerous other examples of the model being able to combine diverse disciplines
13/n

So where there areas where it faltered?
Yes, most certainly. For example, with music. Read for yourself from the paper:
The model failed at any non-trivial form of harmony. As a musician, I find this VERY interesting. Does harmony actually require a higher order intelligence?14/n

GPT-4 was able to carry out coding tasks fairly well. Not just at coming up with code, but also in understanding existing pieces of code and reverse-engineering assembly code.

However the model failed at more complicated mathematical problems, since, being a language model, 15/n

it is context-dependent. The authors look at three aspects of mathematical understanding, viz.,
Creative Reasoning - The ability to identify which arguments, intermediate steps, calculations or algebraic manipulations are likely to be relevant at each stage, 16/n

in order to chart a path towards the solution. - In this, the model demonstrates a high level of ability in choosing the right argument or path towards the solution.

17/n

Technical Proficiency - The ability to perform routine calculations or manipulations that follow a prescribed set of steps. The model falters quite often when doing this making simple arithmetic mistakes.

18/n

Critical Reasoning - Ability to critically examine each step of the argument, break it down into its sub-components, explain what it entails, how it is related to the rest of the argument and why it is correct. The model demonstrates a significant deficiency in this aspect. 19/n

What explains the limitations of GPT-4?
The authors attribute it to two main issues:
"Naive" attention mistakes (remember that attention, that wonderful element of the GPT models that reinforces the learning on a particular aspect in its context) but which act as limitations and

Its "linear thinking" as a next-token prediction machine
Since LLMs are based on the next-character or next-word prediction paradigms, their inherent limitations also impact on their generalizability to AGI.
From the paper:

21/n

Interestingly, the authors borrow from Nobel Laureate Daniel Kahneman's seminal work on the fast and slow modes of thinking. Read for yourself as to what they say:
22/n

The paper has many more examples but suffice to say that while we are still far from AGI, GPT-4 does show that there is a concerted effort to start moving rapidly towards the same.
23/n

The authors having pointed out reasons for the model's less than optimal performance in dealing with specific classes of problems, propose directions for future research. In my opinion it is only a matter of time, not if, we move towards Artificial General Intelligence.

THE END